Principles and methodological considerations

When determining the principles and methodological considerations, competent authorities should decide on the most suitable design and features of the stress test based on its intended objectives.

Scope

Competent authorities should clearly define the coverage in terms of portfolios, sectors, geographies, and activities to ensure a balanced and pragmatic approach that reflects the maturity of available methodologies and data.

For cross-border financial groups, competent authorities should ensure that all relevant entities are included in the assessment.

Based on the objective and the time horizon involved, competent authorities should ensure that the balance between quantitative and qualitative information is adequate for the purpose of the exercise⁷ .

Time horizon

The choice of the time horizon should be aligned with the intended objective of the exercise:

• ▪ When assessing financial resilience to adverse but plausible shocks, a short-term horizon (e.g. up to 5 years) should be used, in line with traditional stress testing practices;
• ▪ When assessing the resilience of financial entities ’ business model and strategy a longer-term horizon (at least 10 years, depending on the business activity and in line with financial entities’ long -term commitments) should be adopted.

Scenario design

Competent authorities should identify and select stress test scenarios based on the objectives of the exercise. They should consider the use of scenarios based on the most recent scientific knowledge and elaborated by widely recognised international or regional organisations⁸ They are also encouraged to leverage sectoral and regional trajectories to provide financial entities with an

7 Given the assumption needed on the scenario calibration and modelling issues, as a general principle, the longer the time horizon, the higher the reliance on more qualitative rather than quantitative information (e.g., no accurate recalculation of capital and solvency metric is expected on a long-term horizon). 8 E.g. for climate risks: the Intergovernmental Panel on Climate Change (IPCC), the Network for Greening the Financial System (NGFS), the International Energy Agency (IEA), the United Nations Environment Programme (UNEP), the Joint Research Center of the EU Commission (EU JRC) or national government or non-government bodies; 9 adequate breakdown of macroeconomic variables, while maintaining the overall internal consistency of the scenarios.

Wherever possible, competent authorities should consider the integration of compound risks, i.e., the additional impact arising from the simultaneous or successive occurrence of multiple shocks (whether ESG-related or broader macroeconomic shocks). Additionally, and where relevant, they should strive to assess second-round effects, meaning the indirect and amplified consequences that result from the initial ESG shocks, either by incorporating them directly into the scenarios or as a complementary analysis.

With the aim of integrating material ESG risks into short-term stress testing exercises, competent authorities should, where feasible and appropriate, build on the scenario design used in traditional stress tests, adapting it to reflect ESG-specific transmission channels and incorporating new variables relevant to ESG risks.

When performing longer-term ESG stress testing, competent authorities should consider several distinct scenarios covering the broad spectrum of possible futures and levels of severity (e.g. including tipping points). The reference scenario may draw on the financial entity’s own central scenario, which can serve as a benchmark to test the impacts of alternative, materially different trajectories on strategy and business model resilience (e.g., via risk-adjusted profitability estimates).

Depending on the sophistication of the approach, scenarios may take the form of simple narratives or more detailed quantitative projections.

Competent authorities may decide to apply the same scenarios to all financial entities for comparability or to tailor scenarios to specific categories of financial entities, considering their risk profiles and business models.

Top-down vs. bottom-up approaches

Competent authorities should assess the merits of a top-down versus bottom-up approach, selecting the most suitable method based on the specific objectives, the maturity of financial entities in ESG stress testing, and the nature of exposures covered by the assessment, taking into account the following considerations:

• ▪ A top-down approach, where competent authorities centrally calculate the impacts of the stress test scenarios, ensures a certain degree of comparability across financial entities and provides competent authorities with stronger control over the process and methodology, reducing the burden for the industry.
• ▪ A bottom-up approach, where financial entities calculate the impact of the stress test scenarios themselves, offers higher level of granularity, allowing to capture idiosyncratic elements specific to some portfolios, counterparties, and exposures. It also fosters greater 10 involvement in the stress testing process and helps enhance financial entities’ capacity building.

When using a bottom-up approach, competent authorities should establish clear methodological frameworks to ensure consistency and level playing field across financial entities, including proportionality considerations regarding the expectations on the sophistication of their calculation of impact.

A hybrid approach could be considered, leveraging the strengths of both methodologies. Competent authorities may decide on the appropriate approach based on data availability and model ownership, taking into account that specific portfolios or activities, such as e.g., underwriting risk, might require more detailed data and modelling than feasible under a top-down approach.

Level of granularity

In deciding on the level of granularity, competent authorities should strike an appropriate balance between simplicity and precision, taking into account data availability, quality and methodological maturity. More granular data might be needed to, for example, appropriately capture activity- or entity-level impacts of transition risk or regional/local impacts of physical risks.

At a minimum, competent authorities should consider the following granularity dimensions:

• a. Portfolio level: Differentiation by asset class (e.g., corporate loans, mortgages, sovereign exposures, equity and corporate bond holdings).
• b. Sectoral level: Classification based on industry sector (e.g., high-carbon industries, energy, real estate, agriculture). For specific sectors a higher granularity may be explored (e.g., electric power, sector impacted by a breakthrough technology, fossil fuel-based industries). For real estate, a distinction between commercial and residential real estate exposures may be explored.
• c. Geographical level: Distinction by region (e.g. NUTS level 3) to assess exposure to location-specific ESG risks, particularly physical risks.
• d. Counterparty level: Granularity by individual obligor or groups of obligors where concentration risks are significant.
• e. Risk category: Separate identification of physical risks (acute and chronic climate hazards), transition risks (policy, technology, and market shifts), and other environmental, social, and governance factors (e.g., biodiversity loss, pollution, social and governance risks).

Competent authorities should decide on the appropriate level of data granularity depending on the type of asset/portfolio and the objective of the exercise. For example, for sovereign exposures country-level could be sufficient, while for lending or underwriting portfolios at least regional granularity up to geolocation data should be explored to assess physical risk.

11

For the analysis of funds, competent authorities may consider, when possible and consistent with the applicable sectoral regulation using a look through approach with regard to holdings of funds.

Balance sheet assumptions

For ESG stress tests, competent authorities should start by using a static balance sheet approach, which is commonly applied in traditional stress tests. However, they may consider the structural impacts of financial entities’ transition plan, where feasible and applicable, depending on the time horizon.

For exercises with medium to long term time horizons, a dynamic balance sheet (with or without constraints) approach could be applied to ensure greater realism, potentially focusing on the main balance sheet items, reflecting changes in portfolio composition (evolving exposures, asset reallocations and risk-mitigating actions in response to ESG-related shocks).

Given the uncertainties surrounding transition pathways, policy developments, and market reactions, competent authorities should consider requiring financial entities to justify their planned responses, ensuring that balance sheet adjustments — such as portfolio reallocation, changes in lending or underwriting policies, or shifts in funding structures — are realistic and consistent with their publicly disclosed transition strategies.

Any management actions recognised by financial entities should be assessed based on their feasibility, timeliness, and potential unintended consequences, avoiding excessive reliance on optimistic assumptions. Competent authorities should ensure that a dynamic balance sheet approach with constraints enhances the realism and forward-looking nature of ESG stress tests through incorporating credible plans and managerial responses, while maintaining minimum safeguards against excessive flexibility in assumptions and enhancing comparability in stress test results.

Sample of financial entities

Competent authorities should define the appropriate sample of financial entities to participate in ESG stress testing and/or ESG scenario analysis, ensuring sufficient sectorial/market coverage, taking into account the specified objectives and, in the case of thematic assessments, the focus of the exercise. The design of the stress test should be adapted, where necessary, to reflect the characteristics of different categories of financial entities.

Competent authorities should also consider conducting targeted exercises on specific subsets of entities to address particular vulnerabilities or evolving ESG risk trends.

12 Proportionality

Competent authorities should apply proportionality principles, allowing for methodological simplifications in relation to the size, nature, and complexity of financial entities’ activities , as well as the materiality of the risks . The proportionality principle should extend to reporting requirements, scenario design, and modelling expectations, ensuring an adequate balance between analytical robustness and practical feasibility.

Given the early-stage of development of ESG stress test modelling and data availability, competent authorities may consider limiting the scope of the analysis, ensuring an adequate cost-benefit balance both for competent authorities and the financial entities.

Model limitations

Competent authorities should recognise that all models have inherent limitations, and this applies especially to evolving ESG models. In the case of bottom-up stress test, to mitigate these limitations, and to enhance the robustness and credibility of ESG stress testing and resilience analysis, they should employ complementary analysis to cross-validate results and ensure their plausibility, use benchmarking techniques, and cross-validate results with alternative methodologies.